Bronchoalveolar Carcinoma: Factors Affecting Survival Richard C. Daly, MD, Victor F. Trastek, MD, Peter C. Pairolero, MD, Paul A. Murtaugh, PhD, Ming-Shyan Huang, MD, Mark S. Allen, MD, and Thomas V. Colby, MD Section of General Thoracic Surgery, - . Section of Biostatistics, and Division of Pathology, Mayo Clinic and Mayo Foundation, Rochester, Minnesota
One hundred thirty-four consecutive patients (65 men and 69 women) underwent pulmonary resection for bronchoalveolar carcinoma. Mean age was 65 years. Lobectomy was done in 100 patients, pneumonectomy in 10, segmentectomy in 5, and wedge excision in 19. Only 10 patients had lymph node metastases (7.5%). The neoplasm was solitary in 111 patients (82.8%); 97 were in stage I, 4 were in stage 11, 9 were in stage IIIa, and 1was in stage IIIb. There were two operative deaths (1.5%). Thirty-nine complications occurred in 31 patients. Median follow-up was 5.1 years. Recurrent bronchoalveolar carcinoma developed in 45 patients. Five- and 10-year survival for patients in stage I was 75.2% and 62.0%, respectively. Survival for patients with T1 NO MO neoplasms was identical to expected survival and was 90.5%
at 5 years, as compared with 55.4% for patients with T2 NO MO disease, only 35.9% for patients with multiple bilateral disease, and 0.0% for patients with bilateral disease ( p < 0.0001). Other significant factors adversely affecting survival included the presence of signs and symptoms, diffuse malignant invasion, mucin-producing tumors, and the histological absence of scar. We conclude that bronchoalveolar carcinoma has a unique natural history that is more influenced by local neoplastic processes than by lymph node metastases. Early aggressive pulmonary resection is safe and offers the potential for cure. The presence of bilateral cancer, however, is ominous.
F
alveolar carcinoma. Six of the 140 patients with BAC were excluded from further analysis because they were first seen with recurrent disease after having had a previous BAC resected elsewhere. The records of the remaining 134 patients were reviewed for presenting signs and symptoms, type of pulmonary resection performed, operative findings, complications, and long-term survival. All chest roentgenograms were reviewed to determine presenting characteristics of BAC. All pulmonary lesions were classified according to the "Glossary of Terms for Thoracic Radiology" as recommended by the Nomenclature Committee of the Fleischner Society [2]. When multiple lesions were present, the largest was used to describe the findings on chest roentgenogram. A pulmonary mass was defined as any distinct lesion 3 cm or greater in diameter. A solitary pulmonary nodule was a welldemarcated lesion less than 3 cm in diameter completely surrounded by pulmonary parenchyma. The lesion was considered an infiltrate if it was a poorly delineated opacity without defined borders that neither destroyed nor displaced gross morphology. Lobar pneumonitis was a diffuse infiltrate involving a minimum of one lobe, often with the presence of an air bronchogram and little or no loss of volume. Each specific lesion was also evaluated for associated cavitation and atelectasis. The microscopic findings of all pathological specimens were reviewed by two of us (M.S.H., T.V.C.). All histo-
irst described in 1876 [l],bronchoalveolar carcinoma (BAC) continues to remain controversial. The numerous names given to this neoplasm over the years (alveolar cell carcinoma, bronchiolar carcinoma, and bronchoalveolar carcinoma) attest to the uncertainties of histogenesis. Although most pathologists today support the existence of BAC as a separate and distinct lung cancer, questions frequently arise as to whether this neoplasm truly differs from conventional pulmonary adenocarcinoma, especially because overlap between features of these two carcinomas does occur and both cancers can coexist in the same patient. Controversy also exists as to the various clinical and pathological findings and as to how these relate to prognosis. Because of this continuing controversy, we have retrospectively reviewed our recent experience with this neoplasm.
Material and Methods Between January 1, 1976, and December 31, 1983, a total of 4,080 consecutive patients with primary lung cancer underwent pulmonary resection at the Mayo Clinic; 1,492 (36.6%)had adenocarcinoma and 140 (3.4%)had bronchoPresented at the Thirty-seventh Annual Meeting of the Southern Thoracic Surgical Association, Dorado, Puerto Rico, November 6-10, 1990. Address reprint requests to Dr Pairolero, Mayo Clinic, 200 First St, SW, Rochester, MN 55905. 0 1991 by The
Society of Thoracic Surgeons
(Ann Thoruc Surg 1991;51:368-77)
0003-4975/91/$3.50
369
DALY ETAL BRONCHOALVEOLAR CARCINOMA
logical assessments were based solely on hematoxylin and eosin stains. Lung carcinomas were accepted as BAC if the neoplasm was an adenocarcinoma that demonstrated a growth pattern along preexisting lung architecture without invasive or destructive growth. Many BACs were associated with localized scarring, and, in such cases, the distinction of stromal invasion from growth along the distorted architectural framework of the scar was arbitrary. In general, the presence of jagged tumor cell nests and a fibroblastic stromal reaction around the cells was considered indicative of stromal invasion. Histologically, each specimen was evaluated for the type of tumor cell (mucinous or nonmucinous) and the presence of localized scarring associated with the tumor. Mucinous cells were defined as those with basally located nuclei and eosinophilic to grayish cytoplasm typical of mucin-producing cells. The following features were also individually assessed: nuclear vacuoles, zones of solid growth within air spaces, papillary growth, clear cell change, necrosis, psammoma bodies, interstitial lymphoid reaction, and interstitial fibrous thickening. The degree of cellular differentiation was graded by Broders’ pathological classification [3]. All solitary neoplasms were postsurgically staged according to the current International TNM Classification System for Lung Cancer [4]; however, if multiple lesions were present, patients were not classified but simply separated into those with unilateral and bilateral disease. Complete lymph node sampling was done in all patients who had curative resection. This included high and low ipsilateral paratracheal, subcarinal, and inferior pulmonary ligament lymph nodes. Any other suspicious lymph nodes were also sampled. Lobar lymph nodes were analyzed in the resected specimens. Operative mortality included all deaths occurring within 30 days of operation and all deaths occurring beyond 30 days but during the same hospitalization. Respiratory failure was defined as mechanical ventilation required for longer than 72 hours, and air leaks were defined as prolonged if they lasted for more than 7 days. In this report, a pulmonary resection was considered curative if all known BAC was removed at the time of initial operation. Recurrent BAC was defined by criteria previously published by us for recurrent lung cancer [5, 61. Survival probabilities were estimated by the KaplanMeier actuarial method [7] using the date of initial operation as the starting date. The Lung Cancer Study Group has recently reported that death is not a suitable therapeutic end-point for the evaluation of lung cancer treatment, as many deaths occur from causes other than lung cancer [8]. Consequently, we have elected to report some of our survival data using death from BAC as the endpoint. However, when comparison of patient survival to expected survival was analyzed, death from all causes was the end-point. Expected survival curves were based on deaths from all causes from west-north central United States 1980 life-table data and were matched for age and sex. The influence of different variables on survival was analyzed with the log-rank test for discrete variables [9] and with the proportional hazards model of Cox for
Table 1 . Preoperative Symptoms No. of Patients
Symptom Cough Sputum production Chest pain Hemoptysis Weight loss Shortness of breath Fatigue Fever Wheezing Osteoarthropathy Herpes zoster
30 23 14 6 6 5 4 4 2 1 1
continuous variables [lo]. Multivariate analysis was used to evaluate all significant univariate relationships using both backward elimination and forward selection processes. Comparison of recurrent BAC and cell types between groups was done with Fisher’s exact test.
Clinical Findings There were 65 male and 69 female patients. Age ranged from 26 to 83 years (median, 65 years). Ninety-one patients (67.9%)were asymptomatic. Cough was present in 30 of 43 symptomatic patients (69.8%).Sputum production and pain were the next most common symptoms (Table 1). Presenting chest roentgenograms demonstrated a single lung lesion in 121 patients (90.3%).Most were seen as an irregular pulmonary mass greater than 3 cm in diameter (78 patients, 58.2%).A solitary pulmonary nod-
Table 2 . Preoperative Chest Roentgenographic Findings Finding Number Single Multiple Location Peripheral Central Description Pulmonary mass >3 cm Solitary pulmonary nodule 5 3 cm Infiltrate Lobar pneumonitis Associated findings Cavitation Atelectasis Diameter 5 3 cm >3 cm
No. of Patients
Percent
121 13
90.3 9.7
132 2
98.5 1.5
78 I5 11 30
58.2 11.2 8.2 22.4
19 7
14.2 5.2
89 45
66.4 33.6
370
DALYETAL BRONCHOALVEOLAR CARCINOMA
Table 3. Other Preoperative Examinations Examination Sputum cytology Bronchoscopy With TBLB Without TBLB Tomograms Computed tomography Transthoracic needle aspiration
Table 5 . Location of Neoplasms
No. of Patients
Additional Information
91 75 49 26
16 32 25 7
74 26
17 12
4
2
TBLB = transbronchoscopic lung biopsy.
ule was uncommon, occurring in only 15 patients (11.2%). Roentgenographic evidence of disease was bilateral in 9 of the 13 patients with multiple lesions (Table 2). In addition to routine chest roentgenograms, sputum cytology, tomograms, and bronchoscopy were the most frequently performed diagnostic tests (Table 3). These tests were considered to have provided additional diagnostic information if they demonstrated additional lesions, evidence of chest wall or mediastinal invasion, lymph node enlargement, or the identification of carcinoma cells. Overall, the diagnosis of carcinoma was made preoperatively in 41 patients (30.6%). Mediastinoscopy was used selectively and was performed in only 15 patients (11.2%);all results were negative. Indication for mediastinoscopy was advanced disease in 11 patients, high risk in 3, and diagnostic in 1. Seven of the 9 patients with bilateral lesions underwent thoracotomy for diagnosis only, 1 had staged bilateral thoracotomies, and 1 had a median sternotomy. Lobectomy was the most frequently performed procedure (70.1%).Wedge excisions were done in 19 patients; only 8, however, had all known cancer removed by this procedure (Table 4). The resection was considered curative in 122 patients (91.0%). Sixty-two patients (46.3%) had tumors involving the right lung, 63 (47.0%)had the left lung involved, and 9 (6.7%) had bilateral lesions (Table 5). At the time of operation, BAC was solitary in 111 patients (82.8%)and multiple in 23 (17.2%).Solitary neoplasms were most
Table 4. Operative Procedure Procedure Pneumonectomy Lobectomy Bilobectomy Segmentectomy Wedge Sternotomy and bilateral wedge Total
No. of Patients 10 94 6 5 18 1 134
Curative Resection 10 93 6 5 7 1
122 (91.0%)
Location Right upper lobe Right middle lobe Right lower lobe Left upper lobe Left lower lobe Multiple unilateral Multiple bilateral
No. of Patients
Percent
28 8 21 37 17 14 9
20.9 6.0 15.7 27.6 12.7 10.4 6.7
commonly located in the upper lobes (58.6%).Eight of the patients with multiple lesions had lesions that numbered greater than ten and all underwent noncurative operations. Synchronous lesions in the remaining 15 patients ranged from two to five (median, two). The distribution by TNM classification is shown in Table 6. Only 10 patients had evidence of lymph node metastases (7.5%);2 patients had N1 disease and 8 had N2. Seventeen of the 23 patients with multiple lesions had no evidence of lymph node metastases; in the remaining 6, lymph node status was unknown because thoracotomy was performed for biopsy only. The parietal pleura was invaded in 6 patients; 3 had T3 lesions, l had T4, and 2 had multiple lesions. Tumor diameter ranged from 3 mm to involvement of an entire lung (median, 2.3 cm). The 30 patients with lobular pneumonitis on presenting chest roentgenogram all had a diffuse infiltrate that involved a minimum of one lobe. At gross examination the involved lobe(s) was consolidated with loss of normal architecture. Histologically, all areas of consolidation were BAC. Twenty of these 30 patients had postsurgical stage I disease (T2 NO
Table 6 . TNM Classification Classification Stage I T1 NO MO T2 NO MO Stage I1 T1 N1 MO T2 N2 MO Stage IIIa T3 NO MO T1 N2 MO T2 N2 MO Stage IIIb T4 NO MO Multiple lesions” Unilateral Bilateral
Number
Noncurative Resection
Operative Death
58 39
0 0
0 1
2 2
0 0
0
3 3 3
1 1 0
0 0 0
1
1
0
14 9
2 7
0
~
a
All sampled lymph nodes were negative for tumor.
0
1
BRONCHOALVEOLAR CARCINOMA
Table 7. Histological Features of Bronchoalueolar Carcinoma Feature
371
DALY ETAL
Ann Thorac Surg 1991;51:360-77
No. Present
Table 9. Location of Recurrent Bronchoalueolar Carcinoma No. of Patients
Location
~
Cell type Nonmucinous Mucinous Mixed
Interstitial findings Thickening Scar Lymphoid reaction Cellular findings Papillae Nuclear vacuoles Solid foci Necrosis Clear cell Gross findings
Diffuse malignant invasion
Contralateral lung 60 55 19 89 72 58 122 75 44 12 6 30
MO); the remaining 10 had multiple tumors. The histological features of all neoplasms are listed in Table 7. Mucinous cell type was more common in the 30 patients with diffuse malignant invasion (66.7%)than in the remaining 104 patients (36.6%;p < 0.01). The neoplasm was grade 1 in 55 patients, grade 2 in 72, grade 3 in 6, and grade 4 in 1.
Results Hospitalization ranged from 4 to 20 days (median, 7 days). Thirty-nine complications (Table 8) occurred in 31 patients (23.1%).There were two operative deaths (mortality, 1.5%).Cause of death was myocardial infarction at 3 days and respiratory failure from adult respiratory distress syndrome at 4 days. Both deaths occurred in patients who had curative resection. Follow-up was complete in 128 of the 132 operative survivors (97.0%) and ranged from 0.5 to 12.8 years (median, 5.1 years). All patients who are currently known to be alive were followed up for at least 5 years (range, 5.0 to 12.8 years; median, 8.1 years). Four patients had
Prolonged air leak Dysrhythmia Atelectasis
Respiratory failure Pulmonary embolus Vocal cord paralysis Myocardial infarction
Ipsilateral lung Bilateral lung
Distant Mediastinal lymph nodes Unknown
1 1
incomplete follow-up; when last evaluated, however, all 4 were still alive and free of BAC, 2 patients at 3 years and 2 at 4 years. Recurrent cancer developed in 45 of the 120 operative survivors who had curative pulmonary resections (37.5%); all recurrences were histologically BAC. New pulmonary lesions (80.0%)were the most common type of recurrence (Table 9) and most were seen as solitary lesions. Time to recurrence ranged from 7 to 73 months with a median of 30 months. Recurrent BAC developed in 10 patients with T1 NO MO disease (17.2%),as compared with 18 with T2 NO MO (47.4%) ( p < 0.05). Bronchoalveolar carcinoma also recurred in 3 patients with stage I1 disease (75.0%),in 5 with stage IIIa (71.4%), and in 12 with multiple neoplasms (unilateral, 90.3%; bilateral, 100.0%). The rate of recurrence in patients with multicentric disease was significantly greater than in patients with T1 NO MO disease ( p < 0.01). Thirty patients with recurrent BAC received chemotherapy, 19 received radiation therapy, and 19 underwent further pulmonary resection, 14 of which were thought to be curative. Overall, 76 patients are known to have died (Table 10). Twelve of these patients had noncurative pulmonary resections and all died of persistent cancer; median survival for these 12 patients was 34 months and ranged from 2 to 46 months. Two other deaths were operative deaths following pulmonary resection. The remaining 62 deaths occurred in the 120 operative survivors who had curative pulmonary resection. Cause of death was recurrent BAC in 44, cardiac in 6, breast cancer in 3, liver failure in 2, renal failure in 1, prostate cancer in 1, colon cancer in 1, trauma in 1, pulmonary embolism in 1, and unknown in
Table 10. Status at Follow-up
Table 8. Postoperative Complications Complication
14 12 10 7
No. of Patients 16 9 8 3 1 1 1
Status
Alive, no disease Alive, with disease Dead, due to BAC Dead, other cause Dead, unknown cause
All Patients
Patients With Recurrence
57 1 56b 18 2
2” 1 41 1
0
a Recurrence developed and was successfully resected with second operIncludes 12 patients who had noncurative resections. ation.
BAC = bronchoalveolar carcinoma.
372
DALYETAL BRONCHOALVEOLAR CARCINOMA
Fig 1 . Overall probability of survival (death from any cause) of 122 patients undergoing curative pulmona y resection for bronchoalveolar carcinoma. Expected survival represents survival of a group of patients matched for age and sex. Zero time on abscissa represents the date of pulmona y resection.
Ann Thorac Surg 1991;51:368-77
100
8
80
-: -
60
3
-
-
BAC (n= 122-
P
One hundred thirty-four consecutive patients (65 men and 69 women) underwent pulmonary resection for bronchoalveolar carcinoma. Mean age was 65 years. Lobectomy was done in 100 patients, pneumonectomy in 10, segmentectomy in 5, and wedge excision in 19. Only 10 patients had lymph node metastases (7.5%). The neoplasm was solitary in 111 patients (82.8%); 97 were in stage I, 4 were in stage 11, 9 were in stage IIIa, and 1was in stage IIIb. There were two operative deaths (1.5%). Thirty-nine complications occurred in 31 patients. Median follow-up was 5.1 years. Recurrent bronchoalveolar carcinoma developed in 45 patients. Five- and 10-year survival for patients in stage I was 75.2% and 62.0%, respectively. Survival for patients with T1 NO MO neoplasms was identical to expected survival and was 90.5%
at 5 years, as compared with 55.4% for patients with T2 NO MO disease, only 35.9% for patients with multiple bilateral disease, and 0.0% for patients with bilateral disease ( p < 0.0001). Other significant factors adversely affecting survival included the presence of signs and symptoms, diffuse malignant invasion, mucin-producing tumors, and the histological absence of scar. We conclude that bronchoalveolar carcinoma has a unique natural history that is more influenced by local neoplastic processes than by lymph node metastases. Early aggressive pulmonary resection is safe and offers the potential for cure. The presence of bilateral cancer, however, is ominous.
F
alveolar carcinoma. Six of the 140 patients with BAC were excluded from further analysis because they were first seen with recurrent disease after having had a previous BAC resected elsewhere. The records of the remaining 134 patients were reviewed for presenting signs and symptoms, type of pulmonary resection performed, operative findings, complications, and long-term survival. All chest roentgenograms were reviewed to determine presenting characteristics of BAC. All pulmonary lesions were classified according to the "Glossary of Terms for Thoracic Radiology" as recommended by the Nomenclature Committee of the Fleischner Society [2]. When multiple lesions were present, the largest was used to describe the findings on chest roentgenogram. A pulmonary mass was defined as any distinct lesion 3 cm or greater in diameter. A solitary pulmonary nodule was a welldemarcated lesion less than 3 cm in diameter completely surrounded by pulmonary parenchyma. The lesion was considered an infiltrate if it was a poorly delineated opacity without defined borders that neither destroyed nor displaced gross morphology. Lobar pneumonitis was a diffuse infiltrate involving a minimum of one lobe, often with the presence of an air bronchogram and little or no loss of volume. Each specific lesion was also evaluated for associated cavitation and atelectasis. The microscopic findings of all pathological specimens were reviewed by two of us (M.S.H., T.V.C.). All histo-
irst described in 1876 [l],bronchoalveolar carcinoma (BAC) continues to remain controversial. The numerous names given to this neoplasm over the years (alveolar cell carcinoma, bronchiolar carcinoma, and bronchoalveolar carcinoma) attest to the uncertainties of histogenesis. Although most pathologists today support the existence of BAC as a separate and distinct lung cancer, questions frequently arise as to whether this neoplasm truly differs from conventional pulmonary adenocarcinoma, especially because overlap between features of these two carcinomas does occur and both cancers can coexist in the same patient. Controversy also exists as to the various clinical and pathological findings and as to how these relate to prognosis. Because of this continuing controversy, we have retrospectively reviewed our recent experience with this neoplasm.
Material and Methods Between January 1, 1976, and December 31, 1983, a total of 4,080 consecutive patients with primary lung cancer underwent pulmonary resection at the Mayo Clinic; 1,492 (36.6%)had adenocarcinoma and 140 (3.4%)had bronchoPresented at the Thirty-seventh Annual Meeting of the Southern Thoracic Surgical Association, Dorado, Puerto Rico, November 6-10, 1990. Address reprint requests to Dr Pairolero, Mayo Clinic, 200 First St, SW, Rochester, MN 55905. 0 1991 by The
Society of Thoracic Surgeons
(Ann Thoruc Surg 1991;51:368-77)
0003-4975/91/$3.50
369
DALY ETAL BRONCHOALVEOLAR CARCINOMA
logical assessments were based solely on hematoxylin and eosin stains. Lung carcinomas were accepted as BAC if the neoplasm was an adenocarcinoma that demonstrated a growth pattern along preexisting lung architecture without invasive or destructive growth. Many BACs were associated with localized scarring, and, in such cases, the distinction of stromal invasion from growth along the distorted architectural framework of the scar was arbitrary. In general, the presence of jagged tumor cell nests and a fibroblastic stromal reaction around the cells was considered indicative of stromal invasion. Histologically, each specimen was evaluated for the type of tumor cell (mucinous or nonmucinous) and the presence of localized scarring associated with the tumor. Mucinous cells were defined as those with basally located nuclei and eosinophilic to grayish cytoplasm typical of mucin-producing cells. The following features were also individually assessed: nuclear vacuoles, zones of solid growth within air spaces, papillary growth, clear cell change, necrosis, psammoma bodies, interstitial lymphoid reaction, and interstitial fibrous thickening. The degree of cellular differentiation was graded by Broders’ pathological classification [3]. All solitary neoplasms were postsurgically staged according to the current International TNM Classification System for Lung Cancer [4]; however, if multiple lesions were present, patients were not classified but simply separated into those with unilateral and bilateral disease. Complete lymph node sampling was done in all patients who had curative resection. This included high and low ipsilateral paratracheal, subcarinal, and inferior pulmonary ligament lymph nodes. Any other suspicious lymph nodes were also sampled. Lobar lymph nodes were analyzed in the resected specimens. Operative mortality included all deaths occurring within 30 days of operation and all deaths occurring beyond 30 days but during the same hospitalization. Respiratory failure was defined as mechanical ventilation required for longer than 72 hours, and air leaks were defined as prolonged if they lasted for more than 7 days. In this report, a pulmonary resection was considered curative if all known BAC was removed at the time of initial operation. Recurrent BAC was defined by criteria previously published by us for recurrent lung cancer [5, 61. Survival probabilities were estimated by the KaplanMeier actuarial method [7] using the date of initial operation as the starting date. The Lung Cancer Study Group has recently reported that death is not a suitable therapeutic end-point for the evaluation of lung cancer treatment, as many deaths occur from causes other than lung cancer [8]. Consequently, we have elected to report some of our survival data using death from BAC as the endpoint. However, when comparison of patient survival to expected survival was analyzed, death from all causes was the end-point. Expected survival curves were based on deaths from all causes from west-north central United States 1980 life-table data and were matched for age and sex. The influence of different variables on survival was analyzed with the log-rank test for discrete variables [9] and with the proportional hazards model of Cox for
Table 1 . Preoperative Symptoms No. of Patients
Symptom Cough Sputum production Chest pain Hemoptysis Weight loss Shortness of breath Fatigue Fever Wheezing Osteoarthropathy Herpes zoster
30 23 14 6 6 5 4 4 2 1 1
continuous variables [lo]. Multivariate analysis was used to evaluate all significant univariate relationships using both backward elimination and forward selection processes. Comparison of recurrent BAC and cell types between groups was done with Fisher’s exact test.
Clinical Findings There were 65 male and 69 female patients. Age ranged from 26 to 83 years (median, 65 years). Ninety-one patients (67.9%)were asymptomatic. Cough was present in 30 of 43 symptomatic patients (69.8%).Sputum production and pain were the next most common symptoms (Table 1). Presenting chest roentgenograms demonstrated a single lung lesion in 121 patients (90.3%).Most were seen as an irregular pulmonary mass greater than 3 cm in diameter (78 patients, 58.2%).A solitary pulmonary nod-
Table 2 . Preoperative Chest Roentgenographic Findings Finding Number Single Multiple Location Peripheral Central Description Pulmonary mass >3 cm Solitary pulmonary nodule 5 3 cm Infiltrate Lobar pneumonitis Associated findings Cavitation Atelectasis Diameter 5 3 cm >3 cm
No. of Patients
Percent
121 13
90.3 9.7
132 2
98.5 1.5
78 I5 11 30
58.2 11.2 8.2 22.4
19 7
14.2 5.2
89 45
66.4 33.6
370
DALYETAL BRONCHOALVEOLAR CARCINOMA
Table 3. Other Preoperative Examinations Examination Sputum cytology Bronchoscopy With TBLB Without TBLB Tomograms Computed tomography Transthoracic needle aspiration
Table 5 . Location of Neoplasms
No. of Patients
Additional Information
91 75 49 26
16 32 25 7
74 26
17 12
4
2
TBLB = transbronchoscopic lung biopsy.
ule was uncommon, occurring in only 15 patients (11.2%). Roentgenographic evidence of disease was bilateral in 9 of the 13 patients with multiple lesions (Table 2). In addition to routine chest roentgenograms, sputum cytology, tomograms, and bronchoscopy were the most frequently performed diagnostic tests (Table 3). These tests were considered to have provided additional diagnostic information if they demonstrated additional lesions, evidence of chest wall or mediastinal invasion, lymph node enlargement, or the identification of carcinoma cells. Overall, the diagnosis of carcinoma was made preoperatively in 41 patients (30.6%). Mediastinoscopy was used selectively and was performed in only 15 patients (11.2%);all results were negative. Indication for mediastinoscopy was advanced disease in 11 patients, high risk in 3, and diagnostic in 1. Seven of the 9 patients with bilateral lesions underwent thoracotomy for diagnosis only, 1 had staged bilateral thoracotomies, and 1 had a median sternotomy. Lobectomy was the most frequently performed procedure (70.1%).Wedge excisions were done in 19 patients; only 8, however, had all known cancer removed by this procedure (Table 4). The resection was considered curative in 122 patients (91.0%). Sixty-two patients (46.3%) had tumors involving the right lung, 63 (47.0%)had the left lung involved, and 9 (6.7%) had bilateral lesions (Table 5). At the time of operation, BAC was solitary in 111 patients (82.8%)and multiple in 23 (17.2%).Solitary neoplasms were most
Table 4. Operative Procedure Procedure Pneumonectomy Lobectomy Bilobectomy Segmentectomy Wedge Sternotomy and bilateral wedge Total
No. of Patients 10 94 6 5 18 1 134
Curative Resection 10 93 6 5 7 1
122 (91.0%)
Location Right upper lobe Right middle lobe Right lower lobe Left upper lobe Left lower lobe Multiple unilateral Multiple bilateral
No. of Patients
Percent
28 8 21 37 17 14 9
20.9 6.0 15.7 27.6 12.7 10.4 6.7
commonly located in the upper lobes (58.6%).Eight of the patients with multiple lesions had lesions that numbered greater than ten and all underwent noncurative operations. Synchronous lesions in the remaining 15 patients ranged from two to five (median, two). The distribution by TNM classification is shown in Table 6. Only 10 patients had evidence of lymph node metastases (7.5%);2 patients had N1 disease and 8 had N2. Seventeen of the 23 patients with multiple lesions had no evidence of lymph node metastases; in the remaining 6, lymph node status was unknown because thoracotomy was performed for biopsy only. The parietal pleura was invaded in 6 patients; 3 had T3 lesions, l had T4, and 2 had multiple lesions. Tumor diameter ranged from 3 mm to involvement of an entire lung (median, 2.3 cm). The 30 patients with lobular pneumonitis on presenting chest roentgenogram all had a diffuse infiltrate that involved a minimum of one lobe. At gross examination the involved lobe(s) was consolidated with loss of normal architecture. Histologically, all areas of consolidation were BAC. Twenty of these 30 patients had postsurgical stage I disease (T2 NO
Table 6 . TNM Classification Classification Stage I T1 NO MO T2 NO MO Stage I1 T1 N1 MO T2 N2 MO Stage IIIa T3 NO MO T1 N2 MO T2 N2 MO Stage IIIb T4 NO MO Multiple lesions” Unilateral Bilateral
Number
Noncurative Resection
Operative Death
58 39
0 0
0 1
2 2
0 0
0
3 3 3
1 1 0
0 0 0
1
1
0
14 9
2 7
0
~
a
All sampled lymph nodes were negative for tumor.
0
1
BRONCHOALVEOLAR CARCINOMA
Table 7. Histological Features of Bronchoalueolar Carcinoma Feature
371
DALY ETAL
Ann Thorac Surg 1991;51:360-77
No. Present
Table 9. Location of Recurrent Bronchoalueolar Carcinoma No. of Patients
Location
~
Cell type Nonmucinous Mucinous Mixed
Interstitial findings Thickening Scar Lymphoid reaction Cellular findings Papillae Nuclear vacuoles Solid foci Necrosis Clear cell Gross findings
Diffuse malignant invasion
Contralateral lung 60 55 19 89 72 58 122 75 44 12 6 30
MO); the remaining 10 had multiple tumors. The histological features of all neoplasms are listed in Table 7. Mucinous cell type was more common in the 30 patients with diffuse malignant invasion (66.7%)than in the remaining 104 patients (36.6%;p < 0.01). The neoplasm was grade 1 in 55 patients, grade 2 in 72, grade 3 in 6, and grade 4 in 1.
Results Hospitalization ranged from 4 to 20 days (median, 7 days). Thirty-nine complications (Table 8) occurred in 31 patients (23.1%).There were two operative deaths (mortality, 1.5%).Cause of death was myocardial infarction at 3 days and respiratory failure from adult respiratory distress syndrome at 4 days. Both deaths occurred in patients who had curative resection. Follow-up was complete in 128 of the 132 operative survivors (97.0%) and ranged from 0.5 to 12.8 years (median, 5.1 years). All patients who are currently known to be alive were followed up for at least 5 years (range, 5.0 to 12.8 years; median, 8.1 years). Four patients had
Prolonged air leak Dysrhythmia Atelectasis
Respiratory failure Pulmonary embolus Vocal cord paralysis Myocardial infarction
Ipsilateral lung Bilateral lung
Distant Mediastinal lymph nodes Unknown
1 1
incomplete follow-up; when last evaluated, however, all 4 were still alive and free of BAC, 2 patients at 3 years and 2 at 4 years. Recurrent cancer developed in 45 of the 120 operative survivors who had curative pulmonary resections (37.5%); all recurrences were histologically BAC. New pulmonary lesions (80.0%)were the most common type of recurrence (Table 9) and most were seen as solitary lesions. Time to recurrence ranged from 7 to 73 months with a median of 30 months. Recurrent BAC developed in 10 patients with T1 NO MO disease (17.2%),as compared with 18 with T2 NO MO (47.4%) ( p < 0.05). Bronchoalveolar carcinoma also recurred in 3 patients with stage I1 disease (75.0%),in 5 with stage IIIa (71.4%), and in 12 with multiple neoplasms (unilateral, 90.3%; bilateral, 100.0%). The rate of recurrence in patients with multicentric disease was significantly greater than in patients with T1 NO MO disease ( p < 0.01). Thirty patients with recurrent BAC received chemotherapy, 19 received radiation therapy, and 19 underwent further pulmonary resection, 14 of which were thought to be curative. Overall, 76 patients are known to have died (Table 10). Twelve of these patients had noncurative pulmonary resections and all died of persistent cancer; median survival for these 12 patients was 34 months and ranged from 2 to 46 months. Two other deaths were operative deaths following pulmonary resection. The remaining 62 deaths occurred in the 120 operative survivors who had curative pulmonary resection. Cause of death was recurrent BAC in 44, cardiac in 6, breast cancer in 3, liver failure in 2, renal failure in 1, prostate cancer in 1, colon cancer in 1, trauma in 1, pulmonary embolism in 1, and unknown in
Table 10. Status at Follow-up
Table 8. Postoperative Complications Complication
14 12 10 7
No. of Patients 16 9 8 3 1 1 1
Status
Alive, no disease Alive, with disease Dead, due to BAC Dead, other cause Dead, unknown cause
All Patients
Patients With Recurrence
57 1 56b 18 2
2” 1 41 1
0
a Recurrence developed and was successfully resected with second operIncludes 12 patients who had noncurative resections. ation.
BAC = bronchoalveolar carcinoma.
372
DALYETAL BRONCHOALVEOLAR CARCINOMA
Fig 1 . Overall probability of survival (death from any cause) of 122 patients undergoing curative pulmona y resection for bronchoalveolar carcinoma. Expected survival represents survival of a group of patients matched for age and sex. Zero time on abscissa represents the date of pulmona y resection.
Ann Thorac Surg 1991;51:368-77
100
8
80
-: -
60
3
-
-
BAC (n= 122-
P
